Date: Apr 15, 2007 11:56 PM
Author: bfinzer@keypress.com
Subject: Re: [Fathom & TP] Fathom as a statistics program

Corey,

We certainly did not design Fathom with the statistical practitioner  
in mind; rather, we designed it for teachers and students for use in  
teaching/learning settings. Lack of built in facility for logistic  
regression, multi-way ANOVA, a multitude of non-parametric tests, and  
a host of other capabilities built into commercial statistical  
packages make Fathom a poor choice for someone doing statistics on a  
regular basis.

On the other hand, we think that the commercial statistical packages  
don't serve the learning environment well at all. The plethora of  
options and output easily overwhelm students (and many teachers) so  
that the focus turns to following the steps rather than understanding  
the statistical inference process.

Have you shown Fathom's simulation capabilities to the prof? Consider  
a setting in which you have a sample of men's and women's incomes  
(perhaps using census microdata), and you want to do a test of  
significance of the difference of median incomes for the two groups.  
How would you do that in a commercial package? But in Fathom it's  
easy to do through scrambling and collecting measures. And each time  
the learner does this kind of simulation they get a hit of the  
fundamental structure of statistical inference, the part that runs  
constant through all the myriad of procedures.

There are surprising settings in which Fathom is being used outside  
the classroom. The environmental protection agency for the state of  
California has a site license for its field workers. Tim Erickson has  
done a series of workshops for a big pharmaceutical company. Many  
school administrators are using Fathom to explore school data. In  
these settings I think it is Fathom's ease of use for exploring data  
that attracts people to it, rather than statistical tests and estimates.

Does the prof think that commercial statistics packages have a place  
in algebra courses? Wouldn't he be in favor of more data appearing in  
mathematics classes? Show him some of the modeling activities that  
might be used in algebra.

Bill

On Apr 13, 2007, at 6:57 PM, Corey Andreasen wrote:

> I've had several conversations with a stat prof out east who has
> little experience with Fathom. He isn't convinced, by what he has
> seen, that Fathom is anything more than a glorified graphing
> calculator. Good for teaching basic concepts, but it certainly isn't
> a program one would use for DOING statistics. He points to an
> inability to do logistic regression as one example of this.
>
> When I see him this summer, what do you think I could show him that
> would convince him otherwise? Any thoughts?
>
> Corey
>
>
> To unsubscribe, send mail to
>      majordomo@mathforum.org
> In the body of the message, use the following command:
>      unsubscribe fathomtp
>
> To reach the (actual, human) owner of the list, send your message to:
>     owner-fathomtp@mathforum.org